Lessons From Red-teaming Auto-review, OpenAI's Blocking Monitor
by monika_j and Giovanni M. Occhipinti
This work and the report were done by Giovanni Occhipinti, as part of the ERA fellowship and mentored by Monika Jotautaitė (Apollo Research). Summary We red-teamed synchronous control measures in Codex, with a focus on its blocking monitor Auto-review. Our campaign directly analyzed the Codex codebase, searching for exploits that...
Oct 3025